Description
Anthropic's mission is to create reliable, interpretable, and steerable AI systems.
The Staff+ Software Engineer, Safeguards Evals role builds the methods and infrastructure for evaluating model behavior and agentic systems.
Responsibilities
- Evaluation research and methodology: Design experiments to improve evaluation quality, develop methods to generate representative test data, and validate grading accuracy.
- Agentic investigation evals: Build and own the evaluation harness for an agentic investigation system, defining metrics, test cases, and grading approaches.
- Datasets grounded in real harm: Construct high-quality eval datasets representing real-world misuse across harm areas.
- Productionization and tooling: Ship successful research into evaluation, regression, and release pipelines, and build tooling for policy experts.
Minimum qualifications
- Experience building and maintaining data pipelines
- Experience working with LLMs and a working understanding of their capabilities and failure modes
- Strong data analysis skills
- Ability to move fluidly between research prototyping and production-quality code
- Ability to translate ambiguous problems into concrete, testable experiments
- Care deeply about AI safety and want your work to have real impact
Preferred qualifications
- 8+ years of industry software engineering or ML engineering experience
- Expertise in building or contributing to LLM or agent evaluation frameworks, benchmarks, or automated grading systems
- Extensive experience in trust and safety, content moderation, or abuse detection systems
- Experience in red teaming, adversarial testing, or jailbreak research on AI systems
The annual compensation range for this role is $320,000-$485,000 USD.
For more fresh new jobs, check out yubhub.co.